Signals Inbox·August 26, 2026·Humanoid Robotics

Will Tesla acquire Figure?

Probably not. Tesla has more reason to watch Figure than it did a year ago, but a near-term acquisition still looks unlikely because Figure is expensive, heavily overlaps with Optimus and may be more useful to Tesla as a rival than as a subsidiary.

Q1Is Tesla actually trying to buy Figure AI right now?

No. As of now, there is no credible public evidence that Tesla is trying to acquire Figure AI.

We found no reported negotiations, takeover approach, investment, partnership or contact between the two companies. Tesla's latest shareholder material talks about building Optimus internally, while Figure's latest disclosures focus on its own customers, manufacturing, Helix AI and data collection.

We should still be careful with the word "no." Tesla disclosed in its latest 10-Q that it acquired an unnamed AI hardware company for $1.95 billion in stock and equity awards during Q2 2026. The target had apparently been bought without a long public rumor cycle. Tesla can keep a technology acquisition quiet.

So a private conversation between Tesla and Figure is possible. There is simply nothing public today showing that such a conversation is happening.

Q2Why does a Tesla-Figure deal look more plausible today than a year ago?

A Tesla-Figure deal looks more plausible today because Figure has become much more credible at the exact moment Tesla has started talking openly about how difficult Optimus is to scale.

Figure has moved beyond occasional humanoid demos. Its April 2026 manufacturing update said BotQ had produced more than 350 Figure 03 robots and cut its demonstrated cycle time from one robot per day to one per hour in less than 120 days. Figure also reported producing more than 9,000 actuators and more than 500 battery packs.

There is real factory evidence too. BMW says Figure 02 worked 10-hour shifts, moved more than 90,000 components and logged roughly 1,250 operating hours while supporting production of more than 30,000 BMW X3 vehicles. Figure 03 has since returned to BMW for a harder logistics workflow, and Figure also signed a commercial agreement with Catalyst Brands to deploy humanoids in distribution and logistics.

Tesla, meanwhile, has become more cautious about Optimus. Its latest shareholder update says the first robots from the new Fremont production line will go into an internal "Optimus Academy" for training-data collection and further development.

A year ago, Tesla buying Figure would have looked like Tesla buying another promising humanoid startup. These days, Figure has enough operating evidence that a shortcut starts to look more tempting.

Q3Is Figure AI actually ahead of Tesla Optimus today?

On the public evidence available today, Figure AI is ahead of Tesla Optimus in autonomous humanoid deployment, while Tesla has the much larger industrial machine behind its robot.

Figure has something unusually valuable in this young market: measurable work outside its own building. BMW independently confirmed the 30,000-vehicle deployment, the 90,000-plus components handled and the 1,250 hours of operation. Figure is now testing a more complicated sequencing workflow with Figure 03.

Its software demonstrations have also become broader. Helix 02 has controlled full-body movement while robots tidy rooms, manipulate deformable objects, operate around furniture and coordinate multiple actions. Company-produced demonstrations are not the same thing as large commercial deployments, but the progression is clear.

Tesla has shown increasingly capable Optimus prototypes and has vastly more experience manufacturing complex machines at scale. What Tesla has yet to disclose is comparable outside-customer humanoid operation. Its latest production plan sends the first Fremont robots into Tesla's own training program.

So if "ahead" means who has shown the stronger public evidence of autonomous humanoids doing useful work today, we give Figure the lead. If it means who could eventually manufacture robots by the hundreds of thousands, Tesla has resources Figure cannot match.

Figure AI vs. Tesla Optimus today

Q4Is Tesla Optimus delayed enough that Tesla might need Figure?

Optimus is delayed enough to make Figure more interesting to Tesla, but Tesla is still investing like a company that expects to solve the problem itself.

Elon Musk gave one of his clearest warnings yet during Tesla's latest earnings call. He described Optimus as the hardest product Tesla has ever had to scale because almost every component is new and there is no mature humanoid supply chain to rely on. He also stressed the difficulty of getting human-level dexterity, reliability and long operating life into the same machine.

Tesla's own production language has become more conservative. The latest shareholder update lists Optimus infrastructure in California and Texas as under construction and gives no installed annual capacity figure. The first Fremont builds are headed to the Optimus Academy rather than customers.

The latest outside check was also less bullish. After visiting Fremont, JPMorgan analyst Rajat Gupta projected commercial Optimus sales in the second half of 2027, later than the timelines investors had previously been discussing.

Still, Tesla has already removed the Model S and Model X manufacturing lines at Fremont and is installing Optimus lines in their place. Construction is also underway in Texas. Those are expensive, physical commitments.

What we see today is a struggling ramp, not a retreat. Figure becomes a serious acquisition candidate only if those problems persist for much longer.

Q5What does Figure have that Tesla cannot easily build itself?

Figure gives Tesla a few things that would take real time to reproduce: a separate humanoid AI stack, experience running robots in other companies' facilities and a rapidly growing human-task data engine.

Helix is the obvious technology asset. Figure built its vision-language-action system specifically around humanoid control. Helix 02 combines perception, reasoning, manipulation and full-body movement rather than treating walking, hands and high-level instructions as separate problems.

Figure has also spent years learning the boring parts of deployment. A robot that works inside another company's factory has to survive shifts, recover from errors, fit existing IT systems, meet safety requirements and maintain useful cycle times. Those lessons are harder to capture in a polished demonstration.

Then there is data. Tesla has an enormous real-world dataset from vehicles, but driving footage contains a very different distribution of actions from folding laundry, moving restaurant equipment, stocking shelves, handling tools or loading parts with human-like hands.

Figure is now attacking that gap much more aggressively. For Tesla, this is probably the most interesting part of the company.

Q6Is Figure's new Index dataset something Tesla would actually want?

Yes. Figure's new Index dataset may be the strongest new reason for Tesla to care about Figure, although we do not think Index alone justifies buying the whole company.

The scale is already unusual. According to Figure's latest Index release, the app has crossed 264,000 downloads in 108 countries, with more than 44,000 weekly active contributors. Users have uploaded more than 16 million videos. Figure says the system is now receiving 30 minutes of video every second, equivalent to 4.9 years of human activity uploaded per day.

The diversity is more interesting than the raw video count. Figure says that every 1,000 hours of accepted Index footage contains 373 distinct tasks, 1,146 manipulated objects and 116 environments. The collection ranges from laundry and cooking to factories, logistics centers, restaurants and retail work.

Figure has already paid contributors $15 million and says it plans to spend more than $1 billion on data and compute over the next 12 months. The pipeline is Figure-exclusive.

This attacks one of the hardest problems in general-purpose robotics: exposing an AI model to enough variations of ordinary physical behavior that the robot can generalize when the room, object, person or task changes.

As seen above, Tesla clearly understands the same problem. Its own Optimus Academy will use early robots for training-data collection and functionality development.

We stop short of calling Index an acquisition-grade moat. Tesla has money, consumer reach, factories, employees and a huge AI organization. If human video at scale becomes essential to Optimus, Tesla can build its own collection network. Index gives Figure a head start. We still need evidence that this head start produces robots that learn materially faster.

Q7Would buying Figure actually speed up Optimus, or just give Tesla two robot programs?

A Figure acquisition would create a huge amount of duplication inside Tesla, and that is probably the strongest strategic argument against the deal.

Tesla and Figure have independently chosen very similar levels of vertical integration. Both are developing humanoid hands, actuators, batteries, perception systems, robot AI, training pipelines, manufacturing processes and dedicated production facilities.

Tesla would immediately face awkward engineering decisions after an acquisition. It would have two robot bodies, two hand designs, two AI stacks, two manufacturing architectures and two internal teams that each believe their system should become the general-purpose platform.

There would certainly be technology worth keeping from both sides. Yet the usual acquisition logic is much cleaner when the buyer is missing one important piece. Tesla's recent $1.95 billion AI hardware purchase fits that pattern far better: acquire specific technology and people, tie a large part of the consideration to service and deployment milestones, then plug the capability into Tesla.

Figure comes as a complete competing system.

Hiring Figure engineers, licensing technology, buying a smaller supplier or copying a good idea is much easier to imagine than Tesla swallowing Figure whole.

Q8Is Figure AI simply too expensive for Tesla to buy?

Figure AI is currently expensive enough that any Tesla acquisition would become a mega-deal rather than a normal technology purchase.

Figure's last disclosed financing valued the company at $39 billion post-money. If we apply a purely illustrative 25% acquisition premium, Tesla would be looking at roughly $48.8 billion.

Tesla reported $43.52 billion in cash, cash equivalents and short-term investments at the end of Q2 2026. A cash purchase around our illustrative takeover price would therefore exceed that entire balance.

And Tesla already has a lot to fund. Its latest 10-Q says 2026 capital expenditure should exceed $25 billion as it spends on AI compute, data centers, manufacturing lines, semiconductors, Optimus and other infrastructure. Q2 alone produced $4.7 billion of operating cash flow against $5.8 billion of capital expenditure, leaving free cash flow at negative $1.1 billion.

Stock makes the arithmetic much easier. Tesla's market value is currently around $1.38 trillion, so a hypothetical $48.8 billion all-stock transaction would equal roughly 3.5% of Tesla's market capitalization.

Tesla could finance Figure. The harder sell would be convincing Tesla shareholders that giving up several percentage points of the company is better than spending a fraction of that amount to keep developing Optimus internally.

Illustrative Tesla-Figure deal math

Q9Has Tesla ever done a deal anything like buying Figure?

Tesla has no acquisition precedent remotely close to the likely size of Figure, even though the company has become more active around AI hardware lately.

Tesla paid roughly $150 million for Grohmann Engineering, an automation specialist. Maxwell Technologies cost about $218 million. Tesla's SEC filings put the accounting purchase price of SolarCity at roughly $2.1 billion. The unnamed AI hardware company bought this year cost $1.95 billion.

The latest AI deal is especially revealing. Tesla paid with common stock and equity awards, with $1.73 billion of the consideration tied to service conditions or performance milestones. That looks like Tesla buying technology and making sure the people responsible for it stay.

A Figure takeover would sit in another category entirely. Even before any premium, its last disclosed valuation is almost twenty times the price Tesla just paid for that AI hardware company.

There is one fresh counterpoint. Musk recently declined to dismiss the possibility of Tesla and SpaceX eventually combining, telling investors that the companies have more and more overlap while noting that any such transaction would need a proper process. Huge corporate combinations are outside Tesla's normal history, though apparently not outside Musk's imagination.

Figure would still require Tesla to make its biggest outside technology bet by a very wide margin.

Q10Would Brett Adcock actually sell Figure to Elon Musk?

Brett Adcock could sell Figure, but Figure's ownership structure and Adcock's recent behavior make a near-term sale difficult to picture.

Forbes currently estimates that Adcock owns around 50% of Figure. The exact private-company cap table is unavailable, and later financing may have changed individual percentages, but he clearly remains unusually important to any transaction.

Figure also keeps tight control over who can own its stock. In a July 2026 notice about unauthorized secondary sales, the company said transfers of both preferred and common shares require board approval and warned that unapproved transfers would be void.

Adcock's old Figure Master Plan also describes a 30-year effort and the ambition to build one of the largest companies in the world. We should not lean too heavily on the document's original claim that Figure was his "sole focus," because that is no longer literally true. Adcock has since started Hark, another AI hardware company.

That fresh development cuts both ways. Business Insider reported this year that Hark raised $700 million at a $6 billion valuation after Adcock initially put $100 million of his own money into the company. Starting another giant project shows that Adcock is willing to reorganize his attention when he sees a new opportunity.

It does not look like preparation to cash out of Figure, though. If anything, Adcock has found a way to build another company while continuing to own an enormous stake in Figure.

Tesla would probably need a very high price and a role that Adcock actually wanted. Money alone is a weaker lever when the founder already owns a stake worth many billions on paper.

Q11Would Figure still be as valuable after Tesla bought it?

Figure would probably lose some commercial flexibility inside Tesla, which cuts into what Tesla would actually be paying for.

An independent Figure can walk into BMW, a retailer, a logistics company or almost any other employer and pitch itself as a neutral robotics supplier. Tesla ownership changes that conversation.

As seen above, BMW is one of Figure's strongest proof points. BMW also competes directly with Tesla in cars. A Tesla-owned robotics supplier operating inside a BMW factory could raise new questions around sensitive production data, factory access and long-term strategic dependence. BMW might still use the robots, but the commercial relationship would become more complicated.

Figure's AI stack creates another integration problem. Helix has been built around Figure hardware and Figure's own training system. Tesla is building Optimus around its own real-world AI organization. Combining the companies would force Tesla to choose what survives and what gets abandoned.

Keeping Figure largely independent would preserve more of its value, but then Tesla would have paid tens of billions for a parallel robot company that still runs its own stack.

This is one reason Figure may be more useful to Tesla as a rival. Figure can discover what works, Tesla can learn from those discoveries, and Tesla keeps full control over Optimus.

Q12What would have to happen before Tesla seriously tries to buy Figure?

Tesla becomes a serious Figure buyer if Optimus keeps slipping while Figure proves that its current lead survives thousands of robots, real customers and long operating hours.

One disappointing Optimus ramp would not be enough. We would want to see repeated delays into 2027 and beyond, persistent reliability or dexterity problems, and a growing gap between Tesla's demonstrations and what other humanoid companies can actually deploy.

Figure would then need to prove the opposite. Hundreds of robots and impressive demos are encouraging. Thousands of robots performing useful work with measurable uptime, low intervention rates and compelling economics would change the calculation much more.

The AI gap would matter too. If Figure can demonstrate that its data engine lets Helix learn new physical tasks much faster than Optimus, Tesla would face a genuine buy-versus-build problem. Today, Figure says its internal generalization results are validating the Index approach, but the company has yet to publish enough evidence for us to quantify that advantage.

Price could finally open a window. A funding downturn, slower valuation growth or a large secondary liquidity need could make Figure easier to approach. Continued technical success combined with another big valuation jump would push a deal further away.

What could change the acquisition probability

Q13So, will Tesla acquire Figure AI?

Probably not. As of today, we would put a Tesla acquisition of Figure AI in the single-digit probability range over the next two years.

Figure has become much more interesting as a potential target. It now has credible factory experience, a fast manufacturing ramp, a serious humanoid AI system and a new way to collect physical-world training data at internet scale.

Tesla also has more reason to pay attention lately. Musk is openly warning investors that Optimus will be extraordinarily hard to manufacture, Tesla's first production robots are still earmarked for internal development, and one recent analyst factory visit pushed expected commercial sales into the second half of 2027.

Yet almost every structural factor still points toward competition.

Tesla has already committed factories, capital, engineers and its broader AI strategy to Optimus. Figure overlaps with that work from the robot's hands all the way up to the AI model. Figure's $39 billion last disclosed valuation makes a takeover huge by Tesla's historical standards. Adcock remains a powerful shareholder with little obvious financial reason to sell. And Figure may actually be commercially more useful as an independent supplier than as a subsidiary of an automaker.

The acquisition thesis becomes much stronger if Tesla eventually concludes that Figure has solved a problem Optimus cannot catch quickly. We do not see that evidence yet.

For now, Tesla and Figure are much more likely to keep racing each other. The strange part is that if Figure keeps winning, Tesla may want it more precisely when it becomes even harder to buy.