The oxford english dictionary added a new meaning for the word snowflake two years ago: a person who is overly sensitive to criticism, or unduly entitled.

After this weeks drama on wall street, it might consider adding another: a promising young tech company with such a hot initial public offering that as its own chief executive suggested wall street investors get over their skis in chasing the share price up.

Whether this spectacular stock market phenomenon will turn out to be transitory, like flakes of frozen water evaporating in the spring sunshine, is just one of the intriguing questions it leaves behind.

The stock market debut of snowflake has certainly gone off with a bang. at a time when wall street is ready to send any cloud software stock with a strong growth story to the stratosphere, the data warehousing company was always guaranteed a warm reception. but a market valuation of about $70bn after the first day of trading or 140 times its current annualised revenues, and nearly six times what the company was judged to be worth in a private fundraising round earlier this year is head-scratching.

The knock-on effects across the entire sector could be powerful and particularly on palantir, whose shares are due to start trading in a direct listing next week. the data analytics company arrives on wall street with plenty of baggage, but it has been remaking itself around a platform called foundry that is not unlike the technology created by snowflake. after years of dithering about whether or not to go public, palantirs timing may turn out to be near-on perfect.

The cloud software companies that have led this years tech stock rally all post impressive growth rates, but it is hard not to see at least some similarities with the dotcom era starting to creep in. one is the infatuation with tam, or total addressable market: the bigger a companys supposed opportunity, the easier it is to forget current nosebleed valuations and look to the horizon.

This is the inevitable result of a platform shift in the computing world. each new generation of computing promises to be much larger than the one preceding it. the cloud, in gestation for more than a decade and a half, has opened the way for services that are instantly consumable: customers no longer need to take on the fixed costs or hassle of building out their own tech infrastructure when they want to fire up a new application.

Frank slootman, the veteran tech chief executive who was brought in to steer the snowflake ipo, said the result has been a flood of contracts that would have seemed impossibly large in earlier generations of software, with million-dollar transactions from companies ive never heard of. snowflake puts its tam at more than $80bn. that implies that its current market share is well below 1 per cent, and leaves the impression of boundless opportunity.

Like the dotcom bubble, however, investors fixated on tams seem to be forgetting one important factor: the competition.

In snowflakes case, this looks about as daunting as it could possibly be. amazon, microsoft and google are fast emerging as the big three among the cloud platforms, the foundations on which the next era of computing is being built (at least, outside china). activities such as holding customers data and providing the analytical tools to make use of it are central to their plans.

That will make snowflake which rides on top of the clouds, helping customers tap into data residing in the underlying platforms an important test of competition in the new era. will the dominant platforms try to exercise their power to squeeze out rivals in strategically important new markets such as this for instance, by trying to bundle analytics with their other services? and will regulators intervene? slacks antitrust complaint against microsoft will be an early test of the regulatory resolve.

As mr slootman pointed out, there are precedents to the position snowflake finds itself in. oracle grew up on ibms computing platform, in a form of symbiosis thatallowed rivals in the database market to also co-operate. also, the giant clouds need to attract vast computing workloads to feed their massive infrastructure, so supporting data-hungry applications such as snowflake has obvious attractions.

But if snowflake owns the ultimate relationship with the customers, that still risks relegating the cloud platforms to the status of commodity infrastructure. it also makes it easier for customers to split their cloud workloads between multiple platforms. that guarantees snowflakes enthusiastic new investors will not get an easy ride.