Ford had the design requirements. Every one of them, written down, in the system. They fed them to the AI and waited for good cars to come out.

The cars came out wrong for three years.

Charles Poon runs vehicle hardware engineering at Ford. On a call with reporters on June 24, he said the company had mistakenly thought that introducing AI and ingesting the design requirements they had would produce a high-quality product. Then he said the part that matters. The AI was “only as good as the information you use to train it,” and in prior years Ford had not paid enough attention to the experience of its most knowledgeable engineers.

The information was there. The requirements were documented. What was missing had never been written down at all.

What was missing was knowing which tolerance looks fine on paper and cracks in the fourth winter, which supplier’s process drifts in August. None of that lives in a requirements document. It lives in someone who has been through enough product cycles to smell it before the numbers say anything.

So Ford went and bought it back. Over three years, 350 veteran engineers. Some were former Ford employees. Others came from suppliers, which is to say from the companies that build Ford’s parts. Ford calls them “gray beards” internally. They now work as internal auditors, running mandatory weekly design reviews to hunt for failure points before a blueprint reaches the plant floor.

It worked. In June, Ford came first among mainstream brands in the JD Power Initial Quality Study for the first time since 2010. Jim Farley told Bloomberg TV the effect on warranty and recall costs adds up to “hundreds and hundreds of millions” of tailwind on cost.

Here is the part nobody is pricing.

Ford did not train those 350 people. It bought them, on a market where they still exist, at a moment when they still exist. Some came off the retirement pile. Others came out of the supply chain, and those suppliers are now building Ford’s parts with that judgment gone from their own floor.

Meanwhile the machine that produces gray beards is being switched off. A Harvard working paper covering 66 million workers across more than 280,000 US firms found entry-level hiring at companies adopting generative AI fell roughly 80% per quarter since 2023, while senior headcount at the same firms kept growing. PwC’s 2026 AI Jobs Barometer found that in the most AI-exposed occupations, 52% of the new skills appearing in junior job ads were skills that used to be asked of experienced hires. In the least exposed, 7%.

Knowledge can be written down, and Ford proved it had been. Judgment gets built by doing the small work badly until you stop doing it badly, and the small work is exactly what is being automated first.

I watched this in the cloud wave. Data center sysadmins renamed themselves cloud engineers, and the workflows did look close enough. Up to a point they were. The concepts carried. Scale did not. The one worth having on your side was the one who had already scaled a platform and eaten the pain of it, and that never showed up on a CV. It showed up the first time something broke in a way the documentation did not cover.

Ford’s escape hatch worked because in 2026 there is still a stock of people who did that work, somewhere, that you can go buy. That stock was made in the 1990s. Nobody is making the 2040 batch.

Every board planning to automate now and hire the veterans back later is counting on the same shelf. It is not being restocked.

The Application

Ford. Over three years it hired 350 veteran engineers, a mix of former employees and people from its suppliers, after AI and automated quality systems failed to deliver. COO Kumar Galhotra said Ford had been relying more and more on automated quality systems and not getting the results it wanted. The veterans now hunt for failure points before parts reach the plant floor. Result: first among mainstream brands in the 2026 JD Power Initial Quality Study, 152 problems per 100 vehicles, 41 fewer than the prior year, first time on top since 2010. Ford has not published how many of the 350 came from suppliers, and no one has asked.

Emma Wiles, Boston University, with three researchers from BCG. Putting AI on the Org Chart. A survey of 1,261 managers found 23% already have AI agents formally listed on their org charts. In the experiment, managers got five documents with planted errors and 20 minutes to review them. The only thing that changed was the label: work from an AI tool, from an AI employee named ALEX-3, or from a human employee named Alex. Among managers whose companies had already put agents on the org chart, the AI employee label produced 18% fewer errors caught, 44% more requests for someone else to review, and a shift of perceived accountability away from the manager and toward the system. Among managers at companies without agents on the org chart, almost no effect. The authors are explicit that this is a governance decision.

Harvard, Hosseini and Lichtinger. Résumé and job posting data on 66 million workers across more than 280,000 US firms, 2015 to 2025. At firms adopting generative AI, entry-level hiring fell roughly 80% per quarter since 2023. Senior employment at those same firms kept rising.

The Noise

“Ford proves AI can’t replace people.”

This is the version that lets a room full of executives exhale, and it is the expensive one.

Ford reversed nothing. The AI quality systems are still installed and the investment continues. The gray beards were brought in to reprogram those systems and mentor junior staff, which means the veterans are being paid to encode the judgment that made them expensive. Ford is still the most recalled automaker in the United States, with 51 recalls this year covering more than 11 million vehicles, more than double the next manufacturer. Galhotra calls recalls a lagging indicator, which may be true and is also what you say when the number is bad.

Then there is the calendar.

On June 27, 2025, at the Aspen Ideas Festival, interviewed on stage by Walter Isaacson, Jim Farley said artificial intelligence was going to “replace literally half of all white-collar workers” in the US. Charles Poon made his confession on June 24, 2026. One year apart, minus three days, same company.

And by the time Farley said it, Ford had already been buying its engineers back for two years. The 350 were hired over three years, which puts the start around 2023. The CEO was forecasting the replacement of white-collar workers in public while his own operation was quietly paying to undo it.

The people his AI could not replace were engineers. White-collar workers. Ford ran the experiment on the exact population Farley said was finished, and had to buy them back at market price.

What Ford says is that its AI could not do this work without the people who knew what right looks like, and that it had let those people go before finding out.

The Question

Take the last thing you automated that is now running in production. Ask who signed off that the output was correct, by name. Who read it and put their name on it.

If the answer is the system, or if nobody can remember, the judgment left your company already and the invoice has not arrived yet. Ford’s arrived three years late and cost hundreds of millions.

Now What?

• Write down the failure modes before you touch the process. Put the people who know what breaks in a room and make them document what they look for, in their words, including the things that sound like superstition. Ford paid for doing this in the wrong order. The window closes when they retire, and you do not get a warning.

• Take the agents off the org chart. Wiles found the oversight damage only shows up where agents are formally institutionalized as employees. Reversing that costs you nothing but a naming convention. Call them tools, because they are tools, and keep the accountability on the person who used them.

• Require a human name on every review of automated output. One person, signed. If nobody will put their name on it, nobody is reading it.

• Look at your suppliers. Ford filled part of its gap out of its own supply chain. Whatever judgment you can buy there, your competitor can buy too, and your suppliers are running the same automation playbook you are. Ask them who reviews their output.

• Count how many juniors joined your area in the last 24 months. That number is your stock of judgment in 2036. If it is zero, you have made a decision about 2036 without discussing it.

What I’m Watching

• Robotics. On July 8, Nature published a UC San Diego preclinical trial where two teleoperated humanoids completed real gallbladder surgeries on large mammals, both in a human-robot pairing and robot-to-robot. Against a dedicated four-arm surgical system, a general-purpose humanoid is versatile, takes less room, and costs a fraction. The question stops being which robot does this surgery and becomes who can operate, and from where.

• Health. On July 1 the FDA approved Casgevy for patients from age two with sickle cell disease or transfusion-dependent beta thalassemia, the first gene therapy cleared for children that young in these conditions. Roughly 5,500 newly eligible patients in the US across more than 75 authorized centers. CRISPR went from its first approval in late 2023 to early pediatric populations in under three years.

• Quantum. The US signed two executive orders on June 22. One pushes commercialization and orders quantum sensors deployed by 2028. The other sets 2031 as the deadline for post-quantum authentication migration. Putting a date on it means the government thinks there is a real chance the hardware exists around then. Microsoft moved its own migration up to 2029, joining Google and Cloudflare.

This is my newsletter. Emerging technology without the hype. Real signals for strategic decisions.

Javier D’Ovidio

For years, the robot that could do real work was always a year away. A clip on a stage. A promise for next year. This year it started coming off a production line.

The most visible ones are shaped like us. Figure, a US robotics company, now builds one of its Figure 03 robots every hour at its California factory, up from one a day in January. 1X started full production of its home robot NEO, with room for 10,000 units a year and a plan for 100,000 by the end of 2027. Boston Dynamics is shipping its Atlas. Agility’s Digit already works, on paid contracts, inside warehouses for Amazon and a Toyota plant in Canada. The hard part everyone said was decades away, building these at scale, got solved fast.

So the wall moves to what comes after the factory: keeping the thing working.

Here is the part that stays out of the launch videos. Most of these machines run on a battery, and under real work that battery does not last a full shift. Agility keeps Digit going by sending it back to a dock to recharge itself. Apptronik’s robot swaps its batteries out. Figure’s charges through its feet. Tesla says its robot can go a full day, a claim the industry is still waiting to see hold up under real load. Every serious builder has the same quiet answer to the same quiet problem. The machine has to stop and get charged, and someone has to design how.

And none of this is about the human shape. A warehouse robot on wheels, a delivery bot, a piece of equipment out in a field, if it runs on a battery, it hits the same wall. The form changes. The constraint does not.

I know this wall from the other side. I am building something physical right now where the battery decides whether it works. Everything clever sits downstream of that one constraint.

We have watched this shape before, one layer up. The AI data center story stopped being about chips a while ago. It became a story about power, about who could get enough electricity to the building. Robots are repeating that move. The next scarce thing is the energy to keep them moving, and the cost of keeping a fleet of them charged.

That is a convergence almost nobody is pricing. Robots and energy are becoming the same problem. A company that ships 10,000 robots a year has built a manufacturing business. Whoever works out how to power and charge 10,000 robots through real working hours has built the business the first one runs on.

The robots are real now. Watch where they plug in.

The Application

Figure. A US robotics company. At its California factory, production of the Figure 03 robot went from one a day in January to one an hour by late April, more than 350 built so far, on the way to 12,000 a year. Forty are working at a BMW plant in South Carolina.

1X. Its California factory began full production of NEO, a home robot, on April 30, with capacity for 10,000 a year and a plan for 100,000 by the end of 2027. It costs $20,000, or $499 a month. The first year of production sold out in five days. First home deliveries are due in 2026, though early demos still ran on a person controlling the robot from a distance.

Agility Robotics. Its robot Digit has the real, named customers: Amazon, the logistics firm GXO, and a Toyota plant in Ontario, Canada, which pay for it by the hour rather than buying it. Digit gets through a shift by parking itself to recharge. The working day is built around the charger, not the robot.

Boston Dynamics. Its Atlas robot has started shipping. All of its 2026 production is already committed to Hyundai and Google’s DeepMind lab, for testing and data, not open sale.

The Noise: “Robots are about to take the jobs.”

Every production milestone this year gets read as a countdown to mass replacement. It makes a good headline and a worse forecast. The robots shipping today do narrow, repetitive work, and they still hand the hard part back to a person. In 1X’s own demo, a human wearing a headset drove the robot through every task. They also stop to charge or swap batteries before a person would take a first break. What arrives in 2026 is a small number of machines doing a few dull jobs under supervision, plugged in for a real part of the day. The replacement story sells conference tickets. The charging schedule is what shows up in the pilot report.

The Question

Take any automation you are being sold that runs on a battery: a robot on your warehouse floor, a machine out in the field, a delivery unit. Ask the seller for the number they leave off the slide, how many hours of real work it does on one charge under your actual load, plus the time it spends charging and the power it draws to keep going as long as you need. Where the math only works with the number on the brochure, you are being sold the demo, not the shift.

Now What?

• Before you sign any robot or battery-powered automation pilot, put the real numbers in the contract, not the demo: hours of work per charge under your actual load, charging time, and how much power a full fleet pulls. Make the seller commit to those.

• Price the energy, not just the robot. Machines that park to charge or swap batteries change your floor layout and your electricity bill. That cost belongs in the business case on day one.

• In any board conversation, separate “you can buy it” from “it works your hours.” A robot in production is not yet a robot doing your shift.

• If you build a physical product that runs on a battery, treat the battery and the charging as the thing that decides everything else. It usually is.

• Watch who moves into powering and charging robots at scale. In computing, the lasting profit went to whoever controlled the electricity. That ground is open here, and mostly empty.

What I’m Watching

• Health. On June 9, Life Biosciences gave the first human dose of a therapy that partly resets cells to a younger state. First time this approach to reversing aging, long the most serious bet in longevity science, has been tried in a living person under an FDA trial.

• Quantum. JPMorgan, quantum firm OQC, and chipmaker AMD opened a shared quantum computing center in London, with the bank as its first dedicated user. A major bank paying for its own access is a different signal than another lab demo.

• Energy. Fervo Energy raised USD 1.89 billion in its stock market debut to build next-generation geothermal power, with Google as its anchor customer for a large block of clean, around-the-clock electricity. Same energy-for-computing pressure now starting to shape robots.

This is my newsletter. Emerging technology without the hype. Real signals for strategic decisions.

Javier D’Ovidio
Exponential Technologist

In May, more than 80% of the code Anthropic shipped to production was written by Claude, not by its engineers. The engineers are still there. They choose the work, review the changes, decide what merges. The thing that writes the code is no longer one of them.

This is the line most organizations are crossing right now, usually without a meeting about it. The line where software stops suggesting and starts doing.

It crossed at Meta, where a support agent built to reset passwords reset them for whoever asked, including the people who took over the Obama White House Instagram account. It crossed on the open web, where since late April most of the traffic moving across it is no longer human. It crossed in Utah, where an AI now renews prescriptions for real patients without a doctor in the loop most of the time.

None of these is the failure most leaders are bracing for. The Meta support agent worked as built. It acted with real permissions, on behalf of an organization that never decided who answers when the output is wrong. That is the actual exposure, and it is already on the network.

I have watched this shape before. Early cloud looked the same. Companies moved their data into the cloud years before they understood the shared-responsibility model, and the breaches that followed came from misconfigured buckets, not from exotic attacks. The capability arrived first. The accountability arrived after the incident. We are in that gap again, this time with software that acts on its own.

The clearest signal is who is worried now. On June 11, Google DeepMind, the lab that put agents at the center of its own products, announced funding for outside researchers to study what happens when millions of agents start interacting with each other. Rohin Shah, who runs their AGI safety work, said the plain part out loud: there isn’t really a field of research for multi-agent safety yet. The people building the actors are asking the rest of us to help work out how they fail together.

The agents are already deployed. The accountability is still a vacancy.

The Application

Anthropic. In its June 4 report “When AI Builds Itself,” the company disclosed that Claude wrote more than 80% of the code merged to production in May 2026, up from low single digits before early 2025. Success on its hardest, least-specified engineering tasks reached 76%, a 50-point jump in six months. The typical engineer now merges eight times as much code per quarter as in the 2021 to 2025 baseline.

Meta. In March, Meta gave its AI support assistant the power to reset passwords and change recovery details across Facebook and Instagram. By June 1, 404 Media reported attackers were taking over high-profile accounts, including the Obama White House and a Space Force senior enlisted account, by asking the agent to link an attacker-controlled email. The bot sent the verification code to the attacker. Meta issued an emergency fix.

Cloudflare. On April 27, automated traffic passed human traffic on the web for the first time. Bots now generate 57.5% of HTTP requests, against 42.5% from people. CEO Matthew Prince had predicted that crossover for 2027.

Doctronic, Utah. STAT reported on May 26 that the state’s AI prescription pilot renews medications in 72% of cases without escalating to a physician. Reviewing doctors later agreed with 91% of those decisions.

The Noise: “The real risk is superintelligence.”

The loudest AI-risk conversation is about a future system smart enough to outwit us. It makes for a gripping keynote. It also aims attention at a threat that hasn’t arrived while the one that has goes unmanaged. We are debating superintelligence while we still cannot solve a requirement as boring as scoping an agent so it doesn’t make decisions in contexts it was never designed for. Meta’s accounts fell to a polite request the support agent was built to honor. No jailbreak required. The present risk is plain: ordinary agents get real permissions and no accountability, and they do what they are asked, including by the wrong people. The version that fills conference halls is years away. The version that fills incident reports is here.

The Question

Walk your operations and mark every place where software now acts instead of suggests: approvals, renewals, account changes, code that ships, messages that send. For each one, name the person who answers when it acts wrongly. The processes where you can’t put a name are your real exposure, and they are easier to find now than during the postmortem.

What I’m Watching

Energy. New small-modular-reactor deals for data centers crossed USD 11.4 billion, while the near-term demand gets covered by gas and coal. Google just signed its first 100 MW “bring your own power” agreement to feed its own load. Today, an AI strategy is partly a fossil-fuel strategy, and the clean replacement is years out.

Robotics. 1X began scale production of its NEO humanoid in Hayward, with capacity for 10,000 units a year. The same automation curve that reached software is now reaching the arm.

Quantum. Microsoft and Quantinuum published an 800x reduction in logical-qubit error rate, peer-reviewed in Nature on June 10. Error correction is the gate every other quantum promise waits behind, and it just moved.

This is my newsletter. Emerging technology without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Javier D’Ovidio

Exponential Technologist

Most enterprise IT leaders called it a curiosity. Too experimental. Too insecure. Not ready for production. Not for serious organizations.

By 2012, the companies that had dismissed it were scrambling to catch up. The ones that had explored it early were running circles around their competitors.

In 2018, machine learning stopped being a research topic and became a product. GPT-2, BERT, the first enterprise AI deployments. Most business leaders called it impressive but premature. Not yet applicable to their industry.

By 2022, the companies without an AI strategy were explaining to their boards why they didn’t have one.

The pattern is consistent. A technology moves from laboratory to infrastructure. A small group moves early. The majority waits for proof. By the time proof arrives, the early movers have compounded their advantage for years.

Quantum is at the beginning of that pattern right now.

The Signal: Thirty Days That Changed the Timeline

In May 2026, three governments made separate decisions within weeks of each other.

The US Department of Commerce proposed $2 billion from the CHIPS Act for quantum hardware and foundries. The first specific quantum allocation in the program, separate from classical semiconductors.

Emmanuel Macron announced an additional €1 billion for France’s quantum plan, bringing total French government investment to over €2.2 billion since 2021.

South Korea expanded its national post-quantum cryptography program to telecommunications, finance, transportation, defense, and space. Real contracts, named companies, defined timelines.

Three governments. Three separate decisions. One month.

When governments move that kind of capital simultaneously, something changed. This is not research funding. This is infrastructure investment with geopolitical stakes. The same signal appeared with semiconductors in 2020, with internet infrastructure in the late 1990s, and with AI regulation in 2023.

The organizations that recognized the pattern early in each of those cycles did not do so because they predicted the future. They did so because they had seen it before.

The Application: What’s Already Running

IBM committed $10 billion to global quantum expansion on May 29, with a stated objective of delivering fault-tolerant quantum supercomputers by 2029. The largest single financial commitment by a technology company to quantum hardware to date. Not a research budget. A delivery commitment with a date.

Oracle and Classiq demonstrated quantum AI agents integrated with Oracle Cloud Infrastructure on May 27. The system generates quantum code from natural language prompts and executes it on OCI resources. They simulated a 12-asset portfolio optimization with 36 qubits, producing results comparable to classical benchmarks. If your organization uses Oracle Cloud, quantum is one API away. No specialized infrastructure. No quantum physicists on staff.

Quantum is not arriving as the science fiction version of itself. It is arriving as a component in hybrid workflows, as a layer of infrastructure security, and as a service integrated into existing cloud platforms.

The Noise: “Quantum Is Still Years Away”

Fault-tolerant quantum computers that run arbitrary algorithms better than classical computers at scale do not exist yet. That version of quantum is years away.

But that is not the only version of quantum that matters for strategic decisions today.

Post-quantum cryptography is a present risk. Data encrypted today can be stolen now and decrypted when quantum computers arrive. South Korea, the US, and Europe are not migrating their cryptographic infrastructure because of a future threat. They are doing it because the window to migrate without urgency is closing now.

Oracle and Classiq demonstrated quantum-as-a-service in production in May 2026.

IBM’s $10 billion commitment has a 2029 delivery date. That is three years from now.

Treating “fault-tolerant universal quantum computing” and “quantum as a strategic variable” as the same thing is how organizations miss the window.

The Question

In 2008, the question was not whether cloud would matter. It was whether your organization would explore it while the cost of learning was low, or scramble to adopt it when the cost of not having it was high.

The same question applies to quantum today.

Which parts of your technology strategy assume that quantum remains irrelevant for the next five years? And what happens to those assumptions if IBM delivers on 2029?

What I’m Watching

The Flatiron-D-Wave debate. The Flatiron Institute published a paper in Science demonstrating a classical algorithm that replicates what D-Wave claimed as quantum advantage. D-Wave responded with a formal rebuttal in arXiv. The debate is unresolved. What matters for business leaders: not all vendor claims of “quantum advantage” have been validated by independent science. The criteria for evaluating quantum vendors is not the same as evaluating conventional software vendors.

Google REPLIQA. Google committed $10 million to a program combining quantum AI and life sciences to simulate biological processes at the molecular level. It is the second signal in two weeks of the quantum-biology convergence, after Annovis Buntanetap’s multi-target approach to Alzheimer’s. The direction is consistent: quantum is entering biology through the door of molecular simulation.

IBM’s timeline pressure on AWS and Azure. If IBM delivers quantum-as-a-service by 2029, Amazon and Microsoft have to respond. That competition benefits every organization already using cloud infrastructure: quantum will arrive in the platform you already use, without requiring a separate decision.

This is my newsletter. Emerging technology without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Javier D’Ovidio

Exponential Technologist

That model is being dismantled in laboratories right now. Not in press releases. In clinical trials.

The Signal: Medicine Finally Asked the Right Question

For most of medical history, the goal was simple: keep people alive longer.

It worked. Global life expectancy went from 47 years in 1900 to over 73 today. But the system that achieved that created a problem nobody planned for: millions of people living longer in worse condition. More years. Fewer of them good.

The United States is the clearest example. Americans spend more per capita on healthcare than any other country. And yet, life expectancy in the US peaked around 2014 and has been declining since. The leading causes of death shifted from infectious disease to chronic conditions: heart disease, diabetes, Alzheimer’s, obesity-related illness. The system got very good at keeping people alive with those conditions. It never figured out how to prevent them at scale.

That’s the distinction between lifespan and healthspan.

Lifespan is how long you live. Healthspan is how many of those years you spend with full physical and cognitive capacity. The gap between the two is where most people spend their final decade, managing chronic disease, losing mobility, watching cognitive function decline.

In the United States alone, chronic disease is projected to cost $47 trillion between 2024 and 2039, including $2.2 trillion annually in medical costs and nearly $900 billion per year in lost productivity by 2039, according to the Partnership to Fight Chronic Disease. 90% of the $4.9 trillion the US spends annually on healthcare goes to people with chronic and mental health conditions, according to the CDC.

The system optimized for the wrong variable for a century.

Jun Axup Penman, longevity researcher and Singularity University expert, put it plainly at the Executive Program in April 2026: nothing proven consistently increases healthspan today. We have lifestyle interventions. We have drugs that manage symptoms. Therapies that reliably extend the years you spend at full capacity don’t exist yet.

That’s the problem the next generation of biotech is working on. And for the first time, the tools available to attack it are fundamentally different.

The roadmap, according to Axup Penman: today, researchers increase healthspan with lifestyle and drugs. In the next five years, replacement and reprogramming therapies enter the picture. The bottleneck isn’t scientific. It’s running clinical trials fast enough to prove what works.

Researchers are using AI to close that gap. A drug discovery process that took a decade now takes 18 months. That compression changes what’s possible in the next five years, not the next twenty.

The Application: What’s Already in the Pipeline

This isn’t future science. These are programs in clinical trials now.

Annovis Bio, Buntanetap. Published Phase 2/3 results in Nature NPJ Dementia in April 2026. A once-daily oral pill targeting amyloid, tau, alpha-synuclein, and TDP-43 simultaneously. Every previous Alzheimer’s drug targeted a single protein. Buntanetap attacks upstream, before the damage forms. Statistically significant results in mild Alzheimer’s patients with confirmed biomarkers. Phase 3 results expected Q1 2027.

Insilico Medicine. First drug designed with AI to reach Phase 2 clinical trials, for idiopathic pulmonary fibrosis. Designed in 18 months. A traditional drug discovery process takes 4 to 6 years to reach the same stage. The company went public in Hong Kong with a $2.7B market cap.

Retro Biosciences, Altos Labs, NewLimit, Calico. Twenty companies working on age reversal through epigenetic reprogramming. Raymond McCauley, founding faculty at Singularity University, frames the science this way: disrupting epigenetics accelerates aging. Repairing it reverses aging. These companies are working on the repair.

The cost of reading the human genome. In 2001, sequencing a human genome cost $100 million. Today it costs under $200. That curve dropped faster than Moore’s Law. At that price, personalized medicine at scale stops being a research concept and becomes an engineering problem.

The market is responding. Longevity biotech captured $3.74 billion in Q1 2026 alone. 56% more than Q1 2025. The sector is projected to grow from $9.86B in 2025 to $29.7B in 2034.

The Noise: Confusing the Two Categories

Lifestyle interventions work. Exercise, sleep, nutrition, stress management. The evidence is solid and the tools are accessible. That’s not noise.

The noise is treating those interventions as a substitute for understanding what’s happening in clinical science.

A CEO who runs marathons and takes magnesium at night isn’t wrong. But if that same CEO dismisses longevity biotech as “wellness stuff,” they’re conflating two categories that operate on completely different timelines, mechanisms, and business implications.

Lifestyle interventions optimize the system you have. What’s happening in labs right now is about reprogramming the system itself.

Those are not the same conversation. And the leaders who treat them as one are going to be surprised by what Phase 3 results look like in 2027.

The Question

Your workforce strategy was designed for people who retire at 65. If healthspan extends by a decade, which assumptions in your talent model, your benefits structure, and your succession planning are no longer accurate?

Now What?

Audit your benefits assumptions. Most corporate health plans were designed for a workforce that peaks at 50 and declines at 60. The economics of those plans change materially if employees work productively at 70. Start the conversation with your CFO and HR lead now, before it becomes urgent.

Map your sector’s exposure. If you operate in pharma, insurance, medical devices, or healthcare-adjacent services, the shift from symptom management to healthspan optimization changes your competitive landscape within five years. The companies building for that world are already funded.

Watch the drug discovery timeline. The 18 months from Insilico is the new benchmark, not the exception. The companies that will lead pharma in 2035 are being built today with AI at the core of their R&D.

Take the Alzheimer’s signal seriously for workforce planning. 55 million people globally live with dementia. Most organizations have no strategy for the caregiving burden this places on their workforce today, let alone what happens if effective treatments arrive and change the cost curve entirely.

What I’m Watching

South Korea executing a national PQC migration. The Ministry of Science and ICT expanded its post-quantum cryptography program to telecoms, finance, transportation, defense, and space. Real contracts, specific companies, defined timelines. When a government moves at this level, it sets the compliance baseline for companies doing business in that country. LATAM doesn’t have an equivalent program. Its trading partners do.

Tesla started Optimus production in Q2 2026. At its Fremont plant, replacing legacy auto lines. The price floor for humanoid robots is being set right now. Every manufacturing cost model that doesn’t include this variable is already outdated.

The genome cost curve keeps falling. At under $200 per genome today, the question for every large employer isn’t whether genomic data will factor into occupational health decisions. It’s who sets the rules when it does.

This is Wavelens. Emerging tech without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Subscribe at newsletter.wavelens.ai so you don’t miss the next edition. Follow Wavelens on LinkedIn for daily signals on the convergence.

Javier D’Ovidio

Wavelens

The robot pays for itself in under two months in the U.S. Under three in Germany. Even in Mexico, the math is closing fast, because the robot’s price drops 20 to 30 percent every year.

It’s happening now.

The Signal: The Price War You’re Not Seeing

The public conversation about humanoid robots is still stuck on two questions: “Will they actually work?” and “Will they take our jobs?”

Both questions were already answered on the factory floors of Guangdong.

A Chinese company called Agibot just produced its 10,000th humanoid robot. The number matters less than the curve: it took them two years to build the first 1,000. One more year to reach 5,000. And just three months to go from 5,000 to 10,000. That kind of acceleration in hardware manufacturing is rare.

Agibot isn’t alone. Unitree plans to ship 20,000 units in 2026. UBTECH targets 5,000 this year and 10,000 next year. In Guangdong, the first fully automated humanoid production line opened in March, producing one robot every 30 minutes, with an annual capacity of 10,000 units. China has 160 humanoid manufacturers, backed by 600 suppliers and 10,000 subcontractors. The official plan: between 28,000 and 100,000 humanoids deployed in factories before the end of 2026.

For context: global humanoid shipments in 2025 reached roughly 13,000 units, according to research firm Omdia. 87% came from Chinese companies. Tesla and Figure AI shipped approximately 150 each.

That’s not a gap. It’s a different game.

Now, why should this matter to a CEO who doesn’t manufacture robots?

Because of what it does to prices. The Unitree G1 sells today for $13,500. A U.S. manufacturing worker costs over $150,000 per year with benefits, overhead, and payroll taxes. That means a G1 pays for itself in under two months of labor savings. And UBTECH says manufacturing costs are dropping 20 to 30 percent annually, with 90% of components already made in China. Industry projections: humanoids below $20,000 by 2026-2027, and in the $10,000 to $15,000 range by 2028-2030.

This is exactly what happened with solar panels, lithium batteries, and drones. China didn’t win the technology race in any of those sectors. It won the production race. And the one who produces at scale sets the price. And the one who sets the price sets the adoption curve.

What happens to your cost structure when your competitor’s humanoids work?

You don’t need to buy a humanoid robot. But your competitor might. Or your supplier. Or a new entrant in your industry that doesn’t carry your legacy labor costs. When a humanoid costs $13,500 and pays for itself in two months, the conversation shifts from “does it work?” to “what happens to my operating costs when someone else adopts it?”

That shift is already underway. And most leaders don’t have it on the radar.

The Application: Already on the Factory Floor

This isn’t a pilot phase.

Renault. 350 Calvin-40 humanoids deploying over 18 months at its Douai plant in France. First brownfield deployment at scale in automotive. Target: 30% reduction in production hours per vehicle. The robots handle tire operations and material transport, tasks that are physically demanding and ergonomically challenging for human workers.

BYD, Geely, FAW-Volkswagen, Foxconn, SF Express. All using UBTECH Walker S2 humanoids in production. Material handling, quality inspection, item sorting. Over $150 million in confirmed orders. These aren’t showcase deployments. They’re running 24-hour shifts.

Toyota. Signed a Robot-as-a-Service agreement with Agility Robotics to deploy Digit robots at its RAV4 plant in Canada. The RaaS model means no capital expenditure upfront, just a monthly fee per robot.

Guangdong. China’s first fully automated humanoid production line, opened March 29. 24 digitalized assembly stages, 77 inspection checkpoints, one robot every 30 minutes.

The contrast: Tesla Optimus has no firm commercial sale date. Figure AI robots cost $150,000 to $200,000. Boston Dynamics Atlas: over $300,000. Western companies are building impressive technology. China is building production capacity.

China is winning the production race. And the production race is the one that sets the price.

The Noise: “Humanoid Robots Are Still a Gimmick”

You’ll hear it after every CES demo. After every Tesla PR event. After every viral video of a robot stumbling on stage.

10,000 robots produced by a single company. 87% of global shipments from one country. Active deployments at BYD, Renault, Toyota, Foxconn. A production line in Guangdong that builds one every 30 minutes.

These are not gimmicks. These are industrial tools in real production environments.

The bias works like this: the West watches CES demos and concludes “it’s not ready yet.” China watches its production numbers and concludes “it already started.”

The demo is over. The deployment started.

The Question: What’s Your Cost Exposure?

If a $13,500 humanoid pays for itself in under two months replacing repetitive physical work, which links in your supply chain, your operations, or your competitor’s operations are vulnerable to that math?

If your team can’t identify those links, your competitor probably already has.

Now What?

Map your exposure. Identify which processes in your organization or supply chain involve repetitive physical work, ergonomically demanding tasks, or continuous shift operations. Those are the first candidates for humanoid automation, not necessarily by your decision, but by your competitor’s or supplier’s.

Benchmark the economics. A humanoid at $13,500 with costs dropping 20 to 30 percent per year. Compare that number against the fully loaded cost of equivalent positions in your operation. If the ROI is under six months, the question shifted from “if” to “when.”

Watch the supply chain bifurcation. U.S. lawmakers have proposed the American Security Robotics Act, a federal procurement ban on Chinese humanoid robots citing national security and data privacy. If it passes, the market splits: Chinese prices keep falling while Western prices stay high. Your sourcing strategy needs to account for both scenarios.

And one more uncomfortable question. If these roles get automated, does your organization have a reskilling plan, or just a cost reduction plan?

What I’m Watching

EY deployed agentic AI to 130,000 auditors in global production. The largest documented enterprise deployment of AI agents. 1.4 trillion lines of accounting data processed annually. Built on Microsoft Azure, Foundry and Fabric. Not a pilot. When the Big Four move into production, the corporate market follows. The enterprise confidence threshold just got crossed.

Big Tech is funding nuclear reactors. Meta signed with Oklo, TerraPower and Vistra for 6.6GW. Microsoft is reactivating Three Mile Island for 2027. Amazon: $700M in X-energy. The same sector that created the energy demand is now forced to finance the supply. They’re venture-capitalizing energy infrastructure.

Cloudflare accelerated its post-quantum deadline to 2029. The second major internet infrastructure player (after Google) to move its PQC timeline forward. A convergent signal with the previous edition: the consensus keeps shifting.

This is Wavelens. Emerging tech without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Subscribe at newsletter.wavelens.ai so you don’t miss the next edition. Follow Wavelens on LinkedIn for daily signals on the convergence.

Javier D’Ovidio Wavelens

THE SIGNAL: The Clock Just Got Shorter

For years, the industry assumed two things about quantum computing.

First, that Q-Day (the day a quantum computer becomes powerful enough to break current encryption) was 10 to 15 years away. Second, that building a quantum computer capable of doing it would require millions of physical qubits.

Both assumptions broke in the last three weeks.

From the top: on March 28, Google announced it moved its internal post-quantum cryptography migration deadline from 2035 to 2029. This isn’t an analyst forecasting. It’s the company with more visibility into quantum hardware than almost any other organization on Earth adjusting its own clock six years forward. Heather Adkins (VP Security Engineering) and Sophie Schmieg explained the decision reflects faster-than-expected progress on three fronts: quantum hardware development, error correction, and factoring resource estimates.

From the bottom: three days later, Caltech and Oratomic published research that reduced the physical qubits needed for a fault-tolerant quantum computer from millions to 10,000-20,000. A 100x reduction in error correction overhead. Dolev Bluvstein, CEO of Oratomic, said: “It is plausible, although not guaranteed, that we will have a fault-tolerant quantum computer by the end of this decade.”

Manuel Endres, one of the co-authors, has already assembled arrays of 6,100 neutral atoms in a lab. The jump from there to 10,000-20,000 is non-trivial, but it’s no longer science fiction.

This isn’t theoretical. In the last two weeks, the general public panicked over Anthropic’s announcement of Claude Mythos, a model capable of finding thousands of zero-day vulnerabilities across every major operating system and web browser. LinkedIn threads. Coverage in Tom’s Hardware, NPR, SecurityWeek. Debates about responsibility. The industry responded more carefully, forming consortia with AWS, Apple, Google, JPMorgan, Microsoft.

And yet: few of the organizations that reacted to Mythos are reacting with the same urgency to quantum. Banks. Pharma. Government. Financial services. Healthcare. Silence.

Mythos finds bugs in your code. Quantum makes your encryption irrelevant. The first is a patching problem. The second is an architecture problem.

And there’s a more uncomfortable difference: the data you’re worried about isn’t safely waiting. Store-now-decrypt-later is an active threat today. Adversaries are capturing encrypted data now, storing it, waiting for the quantum computer to decrypt it later. If your data needs to remain confidential for more than five years (medical records, intellectual property, contracts, regulatory information) you’re already exposed.

Pattern match: the SHA-1 to SHA-2 migration (a much simpler cryptographic transition) took 12 years. The PQC migration is bigger. If Google is right and the real deadline is 2029, that leaves three years to do what the industry took 12 years to do on simpler transitions.

Microsoft targets 2033. NIST plans to deprecate legacy algorithms in 2030, with a final deadline of 2035. NSA: national security systems to PQC by 2027. Most private organizations don’t have PQC migration on their roadmap at all. If they do, it points to 2032-2035. Or it isn’t on the radar.

The gap between what Google sees and what most organizations assume is the blind spot.

This isn’t about quantum. It’s about the gap between what insiders know and what the market assumes. That gap is about to close in an uncomfortable way.

THE APPLICATION: The Infrastructure Already Moving

PQC migration is already underway. Not everywhere.

Google. Chrome already supports post-quantum key exchange. Android 17 integrates ML-DSA (the NIST standard for PQC digital signatures) as a first phase. Google Cloud offers PQC solutions to enterprise customers. Internal deadline: 2029.

Ethereum. Launched pq.ethereum.org this week, a hub dedicated to post-quantum migration. Eight years of accumulated preparation. A concrete roadmap: full migration across four hard forks, targeting 2029. More than 10 client teams shipping weekly devnets.

NSA. CNSA 2.0 requires quantum-safe algorithms for all U.S. national security systems by January 2027.

The contrast is what matters. Microsoft targets 2033, four years behind Google. Bitcoin: official silence. The private sector at large: no mandate.

The organizations moving first share one thing in common: the most to lose if the timeline compresses. Google has Chrome, Android, and Cloud. Ethereum has billions in encrypted value. NSA has secrets that cannot be leaked retroactively. They understand the asymmetric cost of arriving late.

The companies with the most to lose are moving first. The rest assume they have time.

THE NOISE: “Quantum Is Still 10-15 Years Away”

You’ll hear it at conferences. In pitch decks. In vendor keynotes.

It was true two years ago. It isn’t anymore.

Google shortened its deadline to 2029. Caltech demonstrated you need 100x fewer qubits than previously estimated. The CEO of Oratomic, a company founded by researchers from Caltech, Berkeley, Harvard, Amazon, and Google, said “plausible by the end of the decade.”

The problem isn’t whether Q-Day arrives in 2029, 2031, or 2033. The problem is that most organizations don’t have quantum on their radar at all. The ones that do are planning around 2035-2040, and that window has already closed.

And there’s a deeper reason not to trust the old consensus: store-now-decrypt-later is already an active threat. Q-Day’s timing matters less than your exposure timing. If your data needs to stay confidential for more than five years, you’re exposed now.

The timeline didn’t shift. The consensus did.

THE QUESTION: What’s Your Crypto Horizon?

What data in your organization needs to remain confidential beyond 2029?

If that list exists, who is protecting it against something that can be decrypted retroactively?

If your team can’t answer with a specific list in 48 hours, you already have the answer about your state of readiness.

WHAT I’M WATCHING

Insilico Medicine signed a $2.75B deal with Eli Lilly. This closes the loop with the previous edition. In March we said the real bet wasn’t on the drugs but on the AI discovery infrastructure. One week later, Lilly signed the largest AI drug discovery contract to date, including a GLP-1 candidate. It wasn’t a future thesis. It was materializing as we were writing.

Renault plans 350 humanoid robots in 18 months. The first brownfield deployment at scale in the automotive industry. Toyota already signed a RaaS agreement with Agility Digit at its RAV4 plant in Canada. Mind Robotics (a Rivian spinout led by RJ Scaringe) raised $500M. The robotics conversation is shifting from demos to real production deployment.

Google published a 20x reduction in the qubits needed to break ECDLP-256. The same company that moved its PQC deadline also published research showing that breaking blockchain encryption requires fewer than 500,000 physical qubits, 20 times less than previously estimated. A convergent signal: the margins keep compressing in every direction.

This is WaveLens. Emerging tech without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Subscribe at newsletter.wavelens.ai so you don’t miss the next edition. Follow WaveLens on LinkedIn for daily signals on the convergence.

Javier D’Ovidio WaveLens

For thirty years, AI assisted.

It ranked candidates. It flagged anomalies. It suggested the next step. A human decided. A human signed off. A human took responsibility.

That model just changed.

In June 2025, a paper published in Nature Medicine described something that hadn’t happened before: a drug where both the biological target and the molecular compound were identified and designed by AI, completing Phase 2a clinical trials with measurable efficacy results. The drug is called rentosertib. The company is Insilico Medicine. The disease is idiopathic pulmonary fibrosis, a progressive lung condition with no current therapy capable of reversing it.

Traditional drug discovery takes 4 to 6 years to get from target identification to a preclinical candidate. Insilico did it in 18 months. The Phase 2a trial enrolled 71 patients across 21 sites. Patients receiving the highest dose showed a mean improvement in lung function of 98.4 mL. The placebo group declined 20.3 mL. That’s not a rounding error. That’s a direction.

What AI led: finding the target, designing the molecule. What AI didn’t touch: the biology. The trial still took 11 months. Human bodies don’t run on software cycles. That distinction matters, and I’ll come back to it.

But rentosertib isn’t a one-off. Since 2021, Insilico’s platform has nominated more than 20 preclinical candidates. Nine have received IND approval. Multiple Phase 1 trials are active across different disease areas. The model is generating a pipeline at a pace traditional pharma can’t replicate, at a fraction of the cost.

That’s not a drug story. That’s an infrastructure story

We’ve seen this pattern before. When the Human Genome Project completed in 2003, everyone focused on the promise: personalized medicine in five years. Most people missed what was actually happening. The capital wasn’t betting on a specific drug. It was betting on the platform that would find drugs faster. Illumina, founded in 1998, dominated the sequencing market for the next two decades. The leaders who won didn’t move first on the application. They moved first on the layer underneath it.

The capital is making the same bet today. It just doesn’t look the same from the outside.

This has nothing to do with pharma.

AI is shifting from assistant to leader in drug discovery. That shift is already visible in the data. But that paper in Nature Medicine, the one that marks the exact moment, doesn’t come for most industries. The shift happens quietly, in process decisions and vendor pitches and pilot programs that nobody calls historic. And by the time it’s obvious, the infrastructure layer has already been built by someone else.

The Application: The Infrastructure Bet

Insilico proved AI can lead drug discovery. Here’s what happened next.

Eli Lilly inaugurated LillyPod in March 2026: a $1 billion supercomputer built with over 1,000 NVIDIA Blackwell Ultra GPUs and 9,000+ petaflops of capacity. This isn’t a research experiment. It’s a five-year infrastructure commitment to run millions of drug hypotheses in parallel. What Insilico proved with a startup budget, Lilly is building at industrial scale. Even more telling: Lilly is developing TuneLab, a platform that would let other biotech companies access its discovery models. If that scales, the barrier to AI-led discovery drops for the entire sector.

One caveat the field deserves: AI has not improved pharma’s roughly 90% clinical failure rate. It changed the front end, speed and cost of finding candidates. Whether those candidates are better candidates is still an open question.

The platform is the asset. Not the drug.

The Noise: “10 Years to 18 Months”

“AI compresses drug discovery from 10 years to 18 months.”

You’ll hear this in conferences. It’s in pitch decks. It’s technically true and practically misleading.

What AI compressed: the discovery phase. Finding the target, designing the molecule. That compression is real, documented, and significant.

What AI didn’t compress: the biology. The rentosertib Phase 2a trial ran for 11 months, across 71 patients, at 21 sites. Clinical trials take years because that’s how long it takes to observe what a molecule does inside a human being. And most of them still fail.

You can’t iterate a tomato in two-week sprints. You can’t iterate a drug in the human body in 18 months.

The headline collapses two very different things into one clean number. AI improved the front end. The back end follows biology’s clock, not software’s.

The Question: What’s Already Shifting

What technology shifts are already on your team’s radar today, not because you went looking, but because you’re seeing them?

If they answer immediately and without hesitation, you have signal. If they pause, that’s your homework.

What I’m Watching

Quantum computers solving real medical problems this month. The Q4Bio competition (Wellcome Leap) reached its final round with six teams demonstrating that today’s imperfect quantum machines, combined with classical processors, can solve real healthcare problems. Infleqtion is using quantum computing to detect cancer signatures in datasets too large for classical solvers. Algorithmiq is redesigning an oncology drug already in Phase II using quantum-classical hybrid architecture. The narrative says quantum is 10-15 years away. For specific problems in molecular simulation and medical data analysis, partial quantum advantage already exists in 2026. The timeline isn’t uniform, and healthcare is where it’s arriving first.

Humanoid robots crossing the price threshold. Tesla plans 50,000 Optimus units in 2026 at $20,000-30,000 each. Figure AI raised $1 billion and launched Figure 03 for mass manufacturing. The conversation about robots has been about capability for years. The conversation that actually matters is about cost. At these price points, mid-size warehouses and logistics operators can run the numbers. That’s a different wave than “robots in factories.”

Post-quantum cryptography becoming a mandate, not a recommendation. The US government set deadlines for federal agencies to migrate to quantum-safe encryption. Most private sector organizations haven’t started the conversation. The gap between what regulators expect and what organizations have done is widening quietly. If your data needs to stay confidential for the next decade, this is already your problem.

This is Wave Lens. Emerging tech without the hype. Real signals for strategic decisions.

If this edition helped you see something you weren’t seeing before, forward it to one leader in your network who needs to be in this conversation.

Subscribe at wavelens.ai/newsletter so you don’t miss the next edition.

Javier D’Ovidio

Wave Lens

Everyone is talking about AI.

Almost nobody is talking about the fact that AI is just one of five forces converging right now, and that convergence is where the real disruption lives.

Here are the five:

Artificial Intelligence. Already infrastructure, not innovation. The question is no longer whether to use it. It’s understanding what decisions you’re delegating to it without realizing it. Healthcare is adopting AI at twice the rate of the broader economy. AI-designed drugs are entering clinical trials. AI agents are making autonomous decisions in your supply chain. This isn’t coming. It’s here.

Quantum Computing. The industry crossed $1.8 billion in 2025, with projections hitting $125 billion by 2040. Google demonstrated a 13,000x speedup over classical supercomputers. IBM is targeting quantum advantage by 2029. But here’s what most leaders miss: the security implications are already urgent. Every encrypted communication your company has ever sent is potentially vulnerable to “harvest now, decrypt later” attacks. You don’t need to understand qubits. You need to understand what quantum breaks.

Robotics. While the public conversation stays stuck on “will robots take our jobs?”, the real revolution is quieter. 542,000 industrial robots were installed globally in 2024, more than double the number from a decade ago. Humanoid robots are now working in Mercedes factories and Amazon warehouses. Not as experiments. The question isn’t whether robots will transform logistics, manufacturing, and agriculture. It’s whether you’ll notice before your competitors do.

Evolution of Health. AI is reshaping healthcare from reactive treatment to predictive prevention. Wearables are generating continuous health data that edge AI processes in real time, detecting arrhythmias and glucose anomalies before symptoms appear. AI-designed drugs for ALS, autoimmune conditions, and cancer are entering clinical trials. The global AI healthcare market is projected to grow from $26 billion to $187 billion by 2030. This isn’t incremental improvement. It’s a redesign of how medicine works.

Energy & Climate Tech. The energy transition isn’t a policy debate anymore, it’s a technology race. Sodium-ion batteries are entering mass production, offering a cheaper alternative to lithium-ion without the supply chain vulnerabilities. Nuclear fusion achieved regulatory clarity in April 2023 when the US NRC classified fusion reactors differently from traditional nuclear plants, unlocking institutional investment for the first time. Helion Energy already has a power purchase agreement with Microsoft for fusion-generated electricity by 2028. And next-generation nuclear reactor designs are finally breaking free from 20th-century blueprints.

Now here’s the part nobody is telling you.

These five forces don’t operate independently. AI accelerates all four. Quantum will transform AI. Robotics depends on AI. Health and Energy depend on all of the above. The World Economic Forum calls this “technology convergence,” and their research shows that the real value, the real disruption, lives at the intersections.

Leaders who understand AI but ignore quantum are building on infrastructure that may be vulnerable in five years. Leaders who invest in robotics without understanding AI are buying hardware without the brain. Leaders watching energy without understanding the compute demands of AI are missing why data centers now consume as much power as entire nuclear plants.

The signal isn’t any one of these forces.

The signal is the convergence.

And that’s what this newsletter exists to help you navigate.

You just read The Signal, the core of every edition of Wave Lens. Every two weeks, you’ll get five sections. Always the same five. From Edition #1 onward, The Signal analyzes one emerging technology in depth. The Application shows where it’s already working. The Noise calls out what’s overvalued. The Question gives you something to bring to your next meeting. What I’m Watching catches weak signals before they become headlines. That’s it. Consistent. Concise. Useful.

The Application

Where it’s already real.

AI + Healthcare. Google’s DeepMind detected eye diseases in retinal scans as accurately as leading specialists. Microsoft’s Dragon Copilot now listens to clinical consultations and generates notes automatically.

Quantum + Security. The “harvest now, decrypt later” threat is already active. Organizations are collecting encrypted data today with the expectation of decrypting it once quantum computers mature. Large-scale satellite-based quantum key distribution trials are starting in 2026 to protect government communications across Europe and Asia.

Robotics + Logistics. Agility Robotics’ Digit and Apptronik’s Apollo are operating in real warehouse environments, handling totes, bins, and repetitive intralogistics tasks. Not in labs. In production.

Health + AI + Wearables. Closed-loop insulin pumps are autonomously adjusting dosing based on real-time glucose data. Edge AI on wearable devices is detecting cardiac arrhythmias before patients feel symptoms. The shift from reactive to preventive medicine is no longer theoretical.

Energy + AI. AI data centers now require a gigawatt or more of power each, equivalent to an entire conventional nuclear power plant. This energy demand is accelerating investment in next-generation nuclear reactors and fusion research. The two forces are locked in a feedback loop: AI needs energy, and energy needs AI to optimize distribution.

The Noise

“AGI is coming in 2026.”

Every few months, a new headline announces that Artificial General Intelligence is around the corner. It makes for great conference keynotes and generates clicks.

Here’s the reality: we don’t even have a consensus definition of AGI, let alone a timeline. What we do have are increasingly capable narrow AI systems that are transforming specific industries right now.

The danger of the AGI hype isn’t that it’s wrong. It might eventually be right. The danger is that it distracts leaders from the AI decisions that matter today. While you’re debating whether AGI will arrive by 2027 or 2035, your competitors are deploying AI agents that are making autonomous decisions in their supply chains this quarter.

Focus on what AI can do now. Let the philosophers worry about what it might become.

The Question

Which of these five forces will impact our industry first, and what are we doing about it today?

Not next year. Not in our five-year plan. Today.

If your leadership team can’t answer this clearly, you have a strategic blind spot. Bring this question to your next meeting. The conversation it generates will tell you more about your organization’s readiness than any consulting report.

What I’m Watching

• Post-quantum cryptography migration deadlines. NIST finalized its first post-quantum encryption standards in 2024. Organizations have a window to migrate before quantum computers can break current encryption. Most haven’t started. This will become urgent faster than expected.

• Humanoid robot cost curves are dropping faster than expected. If this trajectory holds, affordable humanoid robots could enter mid-market logistics by 2027-2028. Watch the cost per unit, not the demos.

• Fusion energy regulatory clarity. The US NRC classified fusion differently from fission in 2023, and in February 2026 published the proposed regulatory framework. This is the kind of quiet policy shift that unlocks billions in private investment.

Subscribe at wavelens.ai/newsletter/ so you don’t miss the next edition.

Javier D’Ovidio

Exponential Technologist

Wave Lens