When Mike Cannon-Brookes and Scott Farquhar rang the Nasdaq bell in December 2015, it regarded like the beginning of one thing inevitable.
Their firm, Atlassian, wasn’t simply one other software program float, it was a proof level for Australia: an organization in-built Sydney may sit on the identical desk as Silicon Valley’s finest.
Its enterprise mannequin promoting subscriptions to software program regarded like the longer term at a time when manufacturing issues regarded a bit just like the previous.
Ten years later, and the world is in reverse. Atlassian is being faraway from the Nasdaq 100 – a listing of the 100 largest firms on the premier change for tech firms – to get replaced by Sandisk, a {hardware} producer recognized for its reminiscence playing cards and flash drives that lengthy predates the dotcom bubble and the iPhone.
The ejection is a symbolic handover and an illustration of simply how brutally AI is tearing by means of the software program sector.
Sandisk enters with a market cap round $US100 billion ($140 billion); Atlassian exits at roughly $US17 billion. Only a yr in the past, Sandisk shares traded close to $US33. They’re now round $US905 – a 2700 per cent run that might have turned a $35,000 punt into $1 million. Over the identical window, Atlassian has misplaced roughly two-thirds of its worth, with the inventory down 85 per cent from its 2021 peak, regardless of the corporate posting its first billion-dollar cloud quarter.
That’s to not say Atlassian’s enterprise is damaged. By most operational measures, it’s the strongest it has ever been. What’s damaged is the thesis buyers spent a decade shopping for into.
Atlassian was alleged to embody the enterprise software program period which was constructed round recurring income from software program that price virtually nothing to distribute to extra customers. However the market has determined that even sturdy software program franchises might be susceptible if AI adjustments the economics of coding.
The roughest model of the anti-Atlassian thesis is that buyers now worry AI will make its software program simpler to clone, whereas concurrently eliminating most of the software program engineering roles that use it. That’s the reason the inventory has been punished even whereas the corporate has posted sturdy income numbers.
Individuals briefed by Atlassian’s largest clients and backers say the corporate’s bull case is straightforward: it has time. Atlassian’s most essential clients are large ships that take a very long time to show round.
If a significant retailer or monetary establishment desires to alter software program, it will possibly take years and value billions. Simply ask firms such because the ASX or ANZ, which have blown huge sums of cash on expertise initiatives. That’s why when many large firms are contemplating altering their expertise instruments and contemplating distributors’ choices on a scale based mostly on their options, they typically award an incumbent twice as many factors as a newcomer to issue within the issue of switching.
It’s a bonus that might shield Atlassian from being dumped by its largest clients for 18 to 36 months. In the meantime, Atlassian hasn’t been standing nonetheless. It has quietly been shopping for up AI firms reminiscent of Secoda, a little-known agency that helps firms use AI to reply questions from their very own proprietary information.
“Persons are begging Atlassian and [another pre-AI big tech company] Salesforce to catch up,” stated one individual briefed available on the market’s place.
Cannon-Brookes makes a model of the identical argument: final week he referred to as the “vibe-code your personal software program” thesis “nonsense” amongst critical patrons, and identified that solely 40 per cent of Atlassian’s customers are builders anyway.
Sandisk, against this, doesn’t want to inform a narrative. It simply must ship storage. Each information centre coaching a frontier mannequin wants what Nvidia’s Jensen Huang has referred to as “the working reminiscence of the world’s AIs” – storage that AI fashions pull from in actual time.
The semiconductor firm has raised its “worth per bit” by greater than 100 per cent up to now yr and may’t make the stuff quick sufficient.
That sort of progress has introduced opportunists. Final Wednesday, the shoe firm Allbirds, which symbolised the Obama-era Silicon Valley with its wool trainers (actually), accountable enterprise pledges and a $US4 billion public providing – introduced it was promoting itself to a brand-licensing outfit and reinventing the residual shell as “NewBird AI”. Backed by a $US50 million funding, the corporate plans to purchase graphics processing models (GPUs) and lease them to builders.
The inventory, which had cratered earlier than the announcement, jumped 600 per cent in a day. By Thursday, it had already retreated 36 per cent.
When a failing wool runner firm can juice its inventory by hoarding pc chips, we’re firmly within the euphoric part of the {hardware} cycle.
The Atlassian-to-Sandisk swap and the Allbirds-to-NewBird pivot appear to be reverse tales, however they’re truly the identical one. A 40-year-old storage firm and a zombie sneaker model shopping for GPUs now occupy the identical narrative slot that enterprise software program firms held only a yr in the past.
Cannon-Brookes might be proved proper over a five-year horizon, and historical past might look again on this swap as the precise second the market briefly misplaced its thoughts. However, at the least for now, the hierarchy has modified: buyers need both direct AI leverage or a vital infrastructure position. Every part else is yesterday’s information.
The Enterprise Briefing e-newsletter delivers main tales, unique protection and professional opinion. Signal as much as get it each weekday morning.








Leave a Reply