14 DAY TRIAL // The history of Fermi development and mass production is one littered with news, rumors and reports of delays, yield problems and limited availability. Before it finally managed to make its DirectX 11-capable cards available, NVIDA did not only delay them repeatedly, but it even decided to cut down on the number of CUDA cores, because TSMC's 40nm yields were weak. Now, in a surprising move, during the latest conference call, NVIDIA's CEO claimed that, despite all odds, shipments of Fermi graphics adapters numbered in the hundreds of thousands.
Analysts had estimated very poor chip yields, between 20% and 30% to be exact, and this had the consumer base believing that the Santa Clara GPU maker was still having trouble ensuring widespread availability of the GeForce GTX 400 Series and Tesla C2050. Furthermore, the scarcity and high prices of the GTX 470 and GTX 480 did little to discourage this impression. Nevertheless, the company's financial performance during the first quarter seems to support the claims that shipments are going well.
Total revenue for the first three months of fiscal year 2011 (ended May 2) was reported to be of $1.0 billion. This is the equivalent of a 51% on-year increase (from $664.2 million) and a 2% rise compared with the previous fiscal quarter. At the very least, this shows that the graphics-solution developer managed to perform a great deal better than a year ago.
“With our new Fermi-class GPUs in full production, Nvidia's key profit drivers are fully engaged. We shipped a few hundred thousand Fermi processors into strong consumer demand. Our Quadro business for workstations grew strongly, fueled by pent up demand from enterprise customers and new growth markets like video editing. And we had record revenue from Tesla processors for high-end servers,” Jen-Hsun Huang, chief executive officer of Nvidia, said during the latest quarterly conference call with financial analysts.