There is a strange thing happening in private markets right now.
Investors are lining up to buy stakes in companies like OpenAI, Anthropic and SpaceX. But some of the employees who actually own those shares are choosing not to sell.
That matters.
In most private companies, employees are usually happy to take some money off the table when a tender offer comes around, especially when the valuation looks attractive. But the latest behavior from some of the biggest private tech companies suggests employees are thinking differently.
They may believe the next valuation could be much higher.
And when employees who know the business best are unwilling to sell, it creates an interesting problem for investors: there may be plenty of demand for these shares, but very little supply.
OpenAI’s $852B valuation changes the equation
OpenAI is probably the clearest example.
The company’s late-2024 tender offer valued it at around $157 billion.
By March 2026, the valuation attached to its latest funding round had climbed to roughly $852 billion.
That is an enormous jump in a relatively short period.
For employees who could have sold shares at the earlier valuation, the lesson is obvious. Selling can provide liquidity today, but holding can potentially create substantially more value if the company continues to re-rate.
That makes the decision much harder.
If you believe the company could eventually be worth $1 trillion or more, selling today at a lower private-market valuation may feel like giving up too much upside.
And employees are not necessarily making that decision in isolation.
They work inside these companies. They can see the pace of hiring, product development, customer demand and capital spending. They may have a much closer view of the business than an outside investor looking at a term sheet.
That does not mean employees are always right.
But it does mean their willingness, or unwillingness, to sell can tell us something about how scarce the shares have become.
Anthropic is showing the same pattern
Anthropic is seeing similar behavior.
Its most recent tender offer reportedly came up short of the roughly $6 billion of stock investors wanted to purchase.
Employees were reportedly willing to hold at a valuation around $350 billion.
That is an important detail.
Usually, when investors are willing to buy a large amount of stock in a private company, you would expect existing shareholders to take the opportunity to cash out.
Instead, the problem appears to be the opposite.
There are buyers. There just aren’t enough sellers.
That makes private-market pricing more complicated.
If only a small amount of stock is available, the price investors are willing to pay may not simply reflect what they think the entire company is worth.
It may also reflect how difficult it is to get exposure in the first place.
The real story is supply, not just valuation
This is where things get interesting for anyone trying to understand private-company valuations.
A company can have a huge headline valuation, but that number does not necessarily mean there are billions of dollars of shares changing hands at that price.
Sometimes very little stock actually trades.
That creates a supply problem.
Think about it this way:
High demand + limited supply = higher prices.
If investors want exposure to OpenAI, Anthropic or SpaceX but employees are reluctant to sell, buyers have fewer opportunities to get in.
That scarcity can push secondary-market prices higher.
So when you see a private company trading at a particular valuation, it is worth asking a second question:
How much stock actually changed hands to establish that price?
The answer can be just as important as the valuation itself.
SpaceX offers a real-world test
SpaceX gives us an interesting comparison because it has now moved from private-market pricing to public trading.
The company priced its IPO at $135 per share in June, reportedly giving it a valuation of around $1.77 trillion.
Shares then jumped above $200 during the first week of trading.
But they have since fallen below the IPO price.
On Monday, SPCX was trading around $108 to $115.
At first glance, that looks like a simple verdict on the IPO valuation.
It is not quite that straightforward.
SpaceX sold only around 4% of the company in the offering.
That means the public market price is being established by a relatively small portion of the company’s total shares. That price is then used to calculate the value of the entire company.
This is where free float becomes important.
Why free float matters
Free float is the portion of a company’s shares that are actually available for public trading.
It excludes shares held by insiders, locked-up shareholders and other strategic holders who are not freely trading their stock.
SpaceX’s roughly 4% float means that the market is discovering a price using a relatively small pool of shares.
That can create bigger price swings.
If demand rises, the stock can move quickly.
If sentiment changes, it can fall just as quickly.
And the price of that small publicly traded portion does not necessarily tell us what a large private block of shares would be worth in a negotiated transaction.
There is another detail worth remembering.
SpaceX’s December 2025 employee tender was reportedly priced at $421 per share before a five-for-one stock split.
So comparing that old tender price directly with today’s public share price can be misleading without adjusting for the split.
The more defensible takeaway is simpler:
Anyone who bought at the June IPO price is currently underwater, while employees who held through the December tender are not necessarily in the same position.
That distinction matters.
A broker that couldn’t go public is now selling private-market access
There is another interesting development happening alongside all of this.
Clear Street launched Clear Street Private Markets on July 31.
Its first offering gives clients exposure to Databricks, which was valued at around $188 billion in July.
But investors are not simply buying Databricks shares directly.
Instead, they receive an interest in an SPV that holds a stake in a fund that owns the underlying stock.
Clear Street says it could add as many as 30 companies by the end of the year, focusing on private businesses valued between $5 billion and $20 billion.
It has also hired analyst Owen Lau to build dedicated research around private companies.
The timing is notable.
Clear Street withdrew its own $351 million IPO in February after reducing the deal and postponing it.
Now, only months later, the firm is building a business around private-market investing.
That is less a contradiction than a sign of where the firm sees an opportunity.
If companies are staying private longer, there is a growing market for investors who want exposure before an eventual IPO.
Private markets are becoming a bigger business
The traditional path was relatively simple.
A startup raised money privately, eventually went public, and public investors got their opportunity to buy shares.
That model has changed.
Companies are staying private longer.
Employees want liquidity.
Early investors want exits.
New investors want exposure.
And platforms are emerging to connect those groups.
That creates a growing secondary market around private companies.
But there is a catch.
Private-market shares are not the same thing as public-market shares.
They can be harder to value.
Trading can be limited.
Financial information may not be as comprehensive.
And the price of a small transaction can sometimes create a headline valuation that looks much more precise than the underlying market really is.
That is why private-company valuations should be treated carefully.
Financing against private shares adds another layer
Clear Street is also planning to lend against eligible pre-IPO positions.
That could make private-market investing more accessible, but it also introduces another variable.
A buyer using their own cash can potentially sit through a downturn.
A buyer using borrowed money has different constraints.
If the underlying shares fall sharply, the borrower may face pressure to reduce the position or provide additional collateral.
That could matter during a period of stress.
For now, there is not enough public evidence to know exactly how this financing will affect secondary-market pricing during a downturn.
But it is an important part of the market to watch as private-company investing becomes more sophisticated.
The bigger question: are employees right to hold?
This may be the most interesting question in the entire story.
Employees at OpenAI, Anthropic and SpaceX are effectively making a bet.
They are saying that the opportunity cost of selling today is greater than the value of getting liquidity now.
That can make sense when valuations are rising rapidly.
But holding is not risk-free.
Private companies can miss targets.
Valuations can fall.
Liquidity can disappear.
And an employee’s wealth can become heavily concentrated in the company they work for.
So the fact that employees are holding does not automatically mean the companies are undervalued.
It tells us something narrower:
At current prices, many employees appear to believe their shares are worth holding rather than selling.
That is an important signal, but it is not a guarantee.
The Texas power problem could become an AI problem
There is another number worth watching if you’re following AI infrastructure.
474 gigawatts.
That is the reported volume of data-center connection requests sitting in ERCOT’s interconnection queue.
For context, that is more than five times Texas’s record peak electricity demand.
Texas Governor Greg Abbott has ordered state regulators to audit projects in the queue before they move forward, while ERCOT paused its batch-zero review.
The governor’s office said data centers account for around 90% of the requested capacity.
This matters because the AI boom is not only about chips and models anymore.
It is also about electricity.
AI companies can raise enormous amounts of capital and buy huge quantities of computing hardware, but none of that matters if the power and infrastructure are not available to run the data centers.
A regulatory review could therefore become a real timeline risk for AI infrastructure projects.
A few other deals worth watching
Private-market activity is not slowing down.
A few transactions stand out:
-
Valar Atomics: Sequoia led a $1 billion Series B at a reported $6 billion valuation. The nuclear company reportedly tripled its valuation within months after using a fission reaction to power an Nvidia chip.
-
BioCatch: Visa agreed to acquire the Israeli fraud-detection company for $2.4 billion in cash.
-
Horizon3: The autonomous penetration-testing company raised $250 million at a valuation above $2 billion, roughly tripling its previous reported valuation.
-
Olix: The London-based chip company raised $312 million at a $3.3 billion valuation, reportedly the largest funding round for a European chip company.
These deals are useful for understanding where investor appetite is strongest, but funding valuations should not automatically be treated as fair-value estimates.
And then there’s Zepto
One of the more interesting names to watch is India’s quick-commerce company Zepto.
The company delayed its IPO after potential anchor investors reportedly indicated a valuation closer to $2.3 billion, well below the roughly $7 billion private valuation from October 2025.
Instead, Zepto is reportedly looking to raise around $105 million to $120 million at a valuation near $4.5 billion.
That gives us three very different numbers for the same company over a relatively short period.
And that is exactly why private-market pricing deserves scrutiny.
The next question is not just what valuation Zepto gets.
It is who is willing to put new money in at that price.
If outside investors participate, that gives the valuation a different signal than a transaction mainly involving existing shareholders.
What this means for investors
The biggest lesson here is that private-company valuations are becoming harder to interpret.
OpenAI can be worth $852 billion on paper.
Anthropic can command a $350 billion valuation.
SpaceX can reach a $1.77 trillion valuation and then see its publicly traded shares fall below the IPO price.
All of those things can happen at the same time.
The market is not necessarily broken.
It is simply operating with different pools of capital, different levels of liquidity and different amounts of available stock.
The price of a private company is only as strong as the market behind that price.
And when employees refuse to sell, that market can become extremely thin.
For investors, that means the better question is not always:
“What is this company worth?”
Sometimes the better question is:
“Who is actually selling at that price?”
That is where the real signal may be hiding.