We all have seen Getco take a stake in the US broker Knight earlier this Summer, when a trading algorithm put its balance sheet $440m in the red. As part of the cash and share deal announced this week, however, Getco would pay around $540m for the remaining part of the company, and become a public entity. This to us is final straw that puts Getco into the light from the shadow of its non-banking activities. The first was when it started going after client business, which apparently has started taking place since the beginning of this year. What is behind this strategy? Historically Getco has always been trading on its own account with proprietary algorithms, more high than low frequency usually, in any asset class where active listed markets enable them to do so. But liquidity has dried up in many asset classes with the crisis, alongside increased volatility and less predictability of markets, hence their historical revenue streams have become more uncertain. On the other hand, investors are seeking algorithms to help them execute orders more automatically and electronically across asset classes, at a lower cost. Their brokers are not able to provide them with the principal capital that they were used to getting because the Basel III regulations make it now too expensive for their balance sheet, pushing some to develop a different niche around algorithms. Which algorithms are the buy-side going to choose? The ones that are built by a historical prop trader that has built its track-record on algos? Or the ones of their bank? The answer is not that obvious as some banks will keep providing some capital to their good clients through sophisticated internalization matching systems, and as they also provide their clients with access to primary issues. We will certainly keep watching this space.